Research map / Video / Action recognition
Oops! Predicting Unintentional Action in Video
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use video-speed prediction instead of video-context prediction for representation pretraining | Oops! intentionality recognition · Oops! classification, all labels | accuracy: 50.0 -> 53.4 | Helped |
| Use video-speed prediction instead of video-sorting prediction for representation pretraining | Oops! intentionality recognition · Oops! classification, all labels | accuracy: 49.8 -> 53.4 | Helped |
| Use video-speed prediction instead of video-context prediction for representation pretraining | Oops! intentionality recognition · Oops! classification, 10% labels | accuracy: 47.2 -> 49.9 | Helped |
| Use video-speed prediction instead of Kinetics supervision for linear features | Oops! intentionality recognition · Oops! classification, all labels | accuracy: 53.6 vs 53.4 | No clear effect |
| Use Kinetics supervision instead of video-speed pretraining | Oops! intentionality recognition · Oops! classification, fine-tuned with all labels | accuracy: 61.6 -> 64.0 | Helped |
| Use video-speed prediction instead of video-context prediction for pretraining | Oops! temporal localization · within 1 sec | accuracy: 52.0 -> 65.3 | Helped |
| Use video-speed prediction instead of video-context prediction for pretraining | Oops! temporal localization · within 0.25 sec | accuracy: 25.3 -> 36.6 | Helped |
| Use video-speed prediction instead of Kinetics supervision | Oops! anticipation · Oops! anticipation, 1.5 seconds into future | accuracy: 56.7 vs 59.7 | Hurt |
Closest papers by meaning
- Learning Goals From Failure (CVPR 2021 · similarity 0.90)
- Look for the Change: Learning Object States and State-Modifying Actions From Untrimmed Web Videos (CVPR 2022 · similarity 0.88)
- End-To-End Learning of Action Detection From Frame Glimpses in Videos (CVPR 2016 · similarity 0.87)
- Action Recognition From Single Timestamp Supervision in Untrimmed Videos (CVPR 2019 · similarity 0.87)
- SpeedNet: Learning the Speediness in Videos (CVPR 2020 · similarity 0.86)
- No More Shortcuts: Realizing the Potential of Temporal Self-Supervision (AAAI 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.